On 6/12/07, Misha Koshelev mk144210@bcm.edu wrote:
This is necessary (bug not sufficient, see my comment in the bug report that I am about to post) to fix bug 8678.
It is also clearly correct, as checking the type "returned" by RegQueryValueEx when it fails makes no sense.
size = 0;
res = RegQueryValueExW(env, var, NULL, &type, NULL, &size);
- if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) || type != REG_SZ)
+ if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) ||
+ (res == ERROR_SUCCESS && type != REG_SZ))
{
RegCloseKey(env);
return res;
This is wrong. If res == ERROR_SUCCESS, then the first condition evaluates to TRUE and we fall through to exit, but then again, res will never be ERROR_SUCCESS after that particular call. Either way, I have a patch ready to send in that fixes the entire action.
http://www.winehq.org/pipermail/wine-patches/2007-June/040336.html
The third version I sent out (yes I know, I was quite sleepy) _is_ correct though.
And that particular call _will_ return ERROR_SUCCESS and _not_ ERROR_MORE_DATA, see: http://msdn2.microsoft.com/en-us/library/ms724911.aspx
Specifically, a call to RegQueryValueEx with a NULL data pointer and a valid size pointer returns ERROR_SUCCESS if the size pointer is filled, _not_ ERROR_MORE_DATA, which is returned if a valid data pointer is passed.
Misha
________________________________
From: James Hawkins [mailto:truiken@gmail.com] Sent: Wed 6/13/2007 1:22 AM To: wine-devel@winehq.org Cc: Koshelev, Misha Vladislavo Subject: Re: msi: Do not use type value if call to RegQueryValueEx failed.
On 6/12/07, Misha Koshelev mk144210@bcm.edu wrote:
This is necessary (bug not sufficient, see my comment in the bug report that I am about to post) to fix bug 8678.
It is also clearly correct, as checking the type "returned" by RegQueryValueEx when it fails makes no sense.
size = 0;
res = RegQueryValueExW(env, var, NULL, &type, NULL, &size);
- if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) || type != REG_SZ)
+ if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) ||
+ (res == ERROR_SUCCESS && type != REG_SZ))
{
RegCloseKey(env);
return res;
This is wrong. If res == ERROR_SUCCESS, then the first condition evaluates to TRUE and we fall through to exit, but then again, res will never be ERROR_SUCCESS after that particular call. Either way, I have a patch ready to send in that fixes the entire action.
-- James Hawkins
On Tue, 2007-06-12 at 23:22 -0700, James Hawkins wrote:
On 6/12/07, Misha Koshelev mk144210@bcm.edu wrote:
This is necessary (bug not sufficient, see my comment in the bug report that I am about to post) to fix bug 8678.
It is also clearly correct, as checking the type "returned" by RegQueryValueEx when it fails makes no sense.
size = 0; res = RegQueryValueExW(env, var, NULL, &type, NULL, &size);
- if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) ||
type != REG_SZ)
if ((res != ERROR_MORE_DATA && res != ERROR_FILE_NOT_FOUND) ||
(res == ERROR_SUCCESS && type != REG_SZ)){
RegCloseKey(env); return res;This is wrong. If res == ERROR_SUCCESS, then the first condition evaluates to TRUE and we fall through to exit, but then again, res will never be ERROR_SUCCESS after that particular call. Either way, I have a patch ready to send in that fixes the entire action.
--- http://msdn2.microsoft.com/en-us/library/ms724911.aspx If lpData is NULL, and lpcbData is non-NULL, the function returns ERROR_SUCCESS and stores the size of the data, in bytes, in the variable pointed to by lpcbData. This enables an application to determine the best way to allocate a buffer for the value's data. ---
Clearly a rewrite is better than changing one line, but I hope your rewrite checks for ERROR_SUCCESS and not ERROR_MORE_DATA in this case.
Misha
-
James Hawkins -
Koshelev, Misha Vladislavo -
Misha Koshelev